Article Summary
Introduction
A significant body of research finds increasing differences in how partisans view one another, with co-partisans evaluated positively and opposing partisans negatively. One widely cited measure of that growing divide is the level of happiness partisans report when asked how they would feel if their child married a member of the opposite party. This is described as a measure of social distance, which goes beyond mere dislike of opposing partisans and toward active social avoidance.
Klar, Krupnikov, and Ryan, however, contend this measure conflates dislike for opposing partisans with dislike for politics and political parties generally. Respondents aren’t avoiding partisans, per se, but uncomfortable discussions about politics. Additionally, they argue the inclusion of such measures in surveys with other political questions heightens respondents’ attentiveness to political attributes and that such questions often only present the hypothetical in-law’s partisanship without other traits or descriptors. This may have the effect of increasing the perceived importance of partisanship to the in-law’s identity.
How much of revealed dissatisfaction with marrying across party lines, then, is due to dislike for partisanship versus dislike for politics? If more so the latter, as the authors contend, the level of affective polarization in America may be overstated, and Americans may be more willing to tolerate opposing viewpoints than previously thought so long as active disagreement is avoided.
Analytical Approach
Klar, Krupnikov, and Ryan field two waves of a survey experiment ahead of the 2016 general election from January 21 to February 1 (n = 2,030) and from July 19 to August 10 (n = 2,136, 1,428 re-surveyed from first wave). To reduce the perception of the experiments being primarily related to politics, neither asked any political questions or included any political information prior to the treatments.
The primary experimental manipulation had three conditions. In the control group, respondents were asked to rank how happy they would feel if their son or daughter married a Democrat, and then the same question with a potential Republican in-law. In the first treatment condition, respondents were additionally told the potential in-law frequently talks about politics, whereas in the second treatment condition the potential in-law discusses politics only rarely. This allows the authors to determine how much of reported happiness for the marriage is due to partisanship versus frequency of political discussion.
Additionally, the authors independently assign respondents to either a “local candidate” condition or a “national party” condition. In the local candidate condition, the potential in-laws relationship to their party is described as “voted for the local Democratic/Republican candidate”, whereas it is described as “votes for the Democratic/Republican party” in the national party condition. This is meant to distinguish between generalized national-party affect from affect related to any candidate in particular.
Main Findings
Across both studies, respondents reported greater unhappiness with their children marrying out-party members who frequently discuss politics compared to when they rarely discuss politics or no information about frequency is given. The control and “rarely discusses politics” conditions had similar levels of unhappiness.
Disaggregating these results by strength of partisan identity, strong partisans are much unhappier with out-party marriages across all groups. For weak partisans, even when the out-party potential in-law discussed politics frequently, less than 40% reported being unhappy, with much lower unhappiness in the control and rarely talks politics conditions.
Additionally, the authors create a binary measure of affective polarization that takes the value of 1 if the respondent is happy about their child marrying and unhappy marrying an opposing partisan and 0 otherwise. Across both studies, they find very low levels of such affective polarization, with only around 20% being described as such across all conditions. Affective polarization is higher and more responsive to frequency of discussing politics among strong partisans. The authors also report the local candidates treatment lowers affective polarization only in wave 2, speculating this may be because of heightened dislike for supporters of Clinton and Trump before the 2016 election.
Finally, the authors assess the stability of affective polarization between waves, regressing change in polarization on the treatment groups and strength of partisanship (interacted), with demographic controls. This analysis shows stronger partisans becoming more affectively polarized across all treatment groups closer to the election.
Implications
Klar, Krupnikov, and Ryan’s findings imply previous claims about widespread affective polarization, especially as it relates to social distance, may be overstated. While partisans certainly do not like members of the opposing party, at least part of that dislike is due to a dislike of politics writ large. Supplementary analyses confirm even dislike of a respondent’s own party has increased over time. To the extent there is affective polarization in the American public, it is largely concentrated in the strongest partisans. For the majority of Americans who lean toward or weakly identify with a party, dislike for potential spouses from the opposing party seems mostly driven by discomfort discussing politics.
Beyond polarization, this research implies a more general willingness to interact with people with whom one may disagree so long as topics of disagreement aren’t frequently discussed. This complicates measurement of social network disagreement, as it seems visibility of disagreement may be an important moderator.
Questions left unanswered
While the national versus local party treatment effect is discussed for the affective polarization dependent variable, analysis for the effect on out-party and in-party happiness is not presented. It is unclear, then, if the affective polarization found in study 2 is primarily driven by an increase in out-party dislike or an increase of in-party affect. Additionally, the authors motivate their analysis with the notion that potential in-laws are often presented to respondents absent any other characteristics or traits, making their partisan identity seem more important. However, the authors again do just that in their experiment, providing no additional characteristics. It is unclear, then, how partisanship may influence affect net of other characteristics.
Methods and Analysis
Was the study and its analyses pre-registered?: No
Did the study rely on proxy variables to measure polarization?: No
Were standard p-value thresholds used (p<.05 or 95% Confidence Intervals that don’t overlap zero)?: Yes
- Largest p-value presented as significant: 0.05
Were correlational results interpreted with causal language?: No
Limitations / Weaknesses
The authors chose to report findings of their main analysis in two panels: proportion of respondents unhappy with an out-party in-law, and proportion of respondents happy with an in-party in-law. Unfortunately, this makes it difficult to compare levels of happiness across partisan conditions, as the proportions being measured are fundamentally different. One can’t reconstruct one measure from another, since the 5-point scale used for happiness includes a “neither happy nor unhappy” option, so 1 – Proportion Happy does not equal Proportion Unhappy. A number of observable implications are obscured by this decision. For example, if weak partisans are primarily responsive to the level of uncomfortable political discussion, not in- versus out-party status, the average happiness levels should be similar across in- and out-party in-laws with low levels of political engagement and different for those with high levels. This may well be the case, but those results are not presented in the paper.
Open Data & Analyses
Does the article make the replication data publicly available?: No
Does the article make the replication analysis scripts publicly available?: No
Article Citation
Klar, S., Krupnikov, Y., & Ryan, J. B. (2018). Affective polarization or partisan disdain? Public Opinion Quarterly, 82(2), 379–390. https://doi.org/10.1093/poq/nfy014
Bibtex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
@article{10.1093/poq/nfy014,
author = {Klar, Samara and Krupnikov, Yanna and Ryan, John Barry},
title = "{Affective Polarization or Partisan Disdain?: Untangling a Dislike for the Opposing Party from a Dislike of Partisanship}",
journal = {Public Opinion Quarterly},
volume = {82},
number = {2},
pages = {379-390},
year = {2018},
month = {05},
issn = {0033-362X},
doi = {10.1093/poq/nfy014},
url = {https://doi.org/10.1093/poq/nfy014},
eprint = {https://academic.oup.com/poq/article-pdf/82/2/379/25095776/nfy014.pdf},
}